JPS63153821A - Alignment device - Google Patents

Abstract

PURPOSE:To detect an alignment mark positively, and to enable alignment having high accuracy at high speed by mounting a diffraction pattern section to diffract illumination beams toward the outside of a projection optical path in a mark region of a mask and an optical reflection section to reflect illumination beams toward the mark on the wafer. CONSTITUTION:A reticle mark 4 is scanned with beltlike spot beams 30, and a photoelectric device 25 outputs a signal 32a having two peaks in response to one of chromium surfaces 4a. Beltlike spot beams 30 are transmitted through a reticle 1, and changed into beltlike spot beams 30' of a wafer and scan a wafer mark 17, and a photoelectric device 29 outputs a signal 32b. Lastly, the wafer mark 17 is conformed to the center of the reticle mark 4, the center of a central section 4b, thus relatively aligning the reticle 1 and a chip 33 (or the wafer 3).

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an alignment apparatus for an IC exposure apparatus, and more particularly to a through-the-lens (TTL) type alignment apparatus using a projection optical system in a projection exposure apparatus.

Generally, in a projection exposure method using an optical lens of a printing apparatus for manufacturing LSI or VLSI, a mask pattern is projected onto a wafer through an imaging projection lens. In particular, in an apparatus that performs projection exposure by reducing a mask pattern onto a wafer by reducing it to about one-tenth, a single wafer is repeatedly exposed to the same pattern. Therefore, each time each chip is printed, it is desirable to observe the relative position of the mask and wafer through a projection lens (through-the-lens method, hereinafter referred to as TTL) to check the state of alignment. Such a method is called step alignment or each alignment.

In recent years, as the line width of patterns on wafers has become approximately 1 μm, higher precision and higher speed step alignment are required.In order to reduce worker fatigue, automation of alignment, or so-called positioning, is required. Auto alignment has become necessary.

Conventionally, in TTL auto step alignment, a wafer alignment mark image and a reticle alignment mark image are scanned by an ITV camera or photoelectrically converted by slit scanning to detect the relative position of the wafer and reticle. A method has been taken to do so. Figure 1 shows [T
A conventional TTL auto-step alignment device using a V-camera is shown. The IC exposure apparatus prints a pattern on a reticle 1 onto a wafer 3 using a reduction projection lens 2. In order to perform TTL auto step alignment, the reticle alignment mark 4 on the reticle 1 and the wafer alignment mark 5 on the wafer 3 are reimaged on the light receiving surface of the ITV camera 14. However, this wafer alignment mark 5 is imaged onto the reticle 1 by the projection lens 2.

The image on the light receiving surface of the ITV camera 14 is given by a circular enlarged portion 17 on the left side of the figure, where image 4' is a reticle alignment mark image and image 5' is a wafer alignment mark image.
It is a statue of Mark. The relative positions of reticle 1 and wafer 3 are detected by photoelectric signals obtained by scanning these images with TV scanning lines, and by moving the position of reticle 1 or the position of wafer 3, reticle 1 and wafer 3 ( (chip) to a determined relative position. Light source 7, leading waveguide 8, condensing lens 9, field stop 10, half mirror 11
, alignment objective lens 12, and mirror 13 constitute an illumination system. Such prior art devices have the disadvantage that the alignment marks have a low contrast, making them difficult to measure. For example, S
t SiO! on the wafer! When detecting a mark formed by Si and SiO□, the contrast between the mark and the wafer is low because the total reflectance of Si and SiO□ is almost equal. Since the photoelectric signal from ITV is a gray level signal depending on its contrast, it may be difficult to detect marks. Furthermore, when alignment is performed using a signal with a low S/N ratio, it is necessary to integrate the signal for a long time in order to improve the alignment accuracy, which reduces the alignment speed. Further, there is also a drawback that the shape of the signal is not uniform due to the pattern formation condition of the alignment mark 5 on the wafer and the influence of resist coating. For example, due to the influence of resist coating, resist interference, etc., the alignment mark may become a dark line or a bright line on the imaging plane.
Alternatively, a phenomenon occurs in which the brightness and darkness are reversed at the edge of the mark. Therefore, mark i1 identification and mark position detection are as follows:
It had become very complicated.

For this reason, attempts have been made to adopt a dark field detection method using a laser beam as illumination light instead of bright field observation using ITV. In the case of dark-field detection, since the scattered and diffracted light generated at the step edge of the mark is detected, the signal is obtained only from the mark while there is no signal around the mark, so the S/N of the signal is low. The ratio is good (it has the advantage of a high detection rate), and dark field detection is said to be less affected by resist coating than bright field detection.

By the way, in either the bright-field alignment optical system or the dark-field alignment optical system shown in Fig. 1, there is one disadvantage when considered as an exposure device.Normally, the reticle alignment mark is the circuit to be exposed. Since it is provided very close to the periphery of the pattern area, when the TTL method 7 alignment is completed, the -1 part of the alignment optical system (for example, the mirror 13 in FIG. 1) must be evacuated outside the exposure area. Such a configuration is disadvantageous in terms of the throughput of the exposure position vjt, and also has the problem that the setting accuracy of the movable optical material must be relatively high.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide an apparatus that can reliably detect alignment marks with a simple configuration and perform high-speed, highly accurate positioning without reducing throughput.

In order to achieve this objective, the present invention provides coherent illumination light ( irradiation means (18, 19, 20, 21, 22, 23) for irradiating laser beam);
, 24, 48, 49), and the mark area of the mask is provided with a diffraction pattern part (60a) that diffracts the illumination light toward the outside of the projection optical path, and a diffraction pattern part (60a) that directs the illumination light to the target object (60a) through the projection optical system.
The light reflecting portion (6) reflects the light toward the mark 17 on the
0b).

Next, specific embodiments of the present invention will be described in detail.
Before that, I will explain the basic technology of the TTL alignment system using a laser beam.

In general, the projection lens of an IC projection exposure apparatus removes chromatic aberration, spherical aberration, etc. only for 1.2 specific wavelengths in order to obtain high resolution. This specific wavelength becomes the wavelength used for exposure. Therefore, light with a wavelength that is not affected by chromatic aberration, spherical aberration, etc. of this projection lens is used as alignment illumination light. Below is an explanation of the basic technology.
And in a particular implementation of the invention, the g-line of a mercury lamp is used as the exposure light source. Therefore, a projection lens with an aberration removed by the g-line is used, and He-Cd (helium cadmium) is used as the illumination light source for alignment mark detection.
Adopt laser. The oscillation wavelength of the He-Cd laser is 441.6 nm, which is much lower than the wavelength of the G-line, which is 436 nm, so that the influence of aberrations can be removed. By using laser light as an illumination light source for alignment mark detection, a long and narrow spot light (band beam) in an arbitrary shape, in this case a slit shape, can be created as scanning light on the reticle surface and the wafer surface. It can be formed to a high degree of accuracy.

2 and 3 show the optical system of the alignment device according to the basic technology of the present invention. In each of FIGS. 2 and 3, the upper square is the reticle l viewed from above, and the lower square is the This is a top view of chips 33 on wafer 3. Moreover, FIG. 3 is a view of FIG. 2 viewed from the right side. A G-line exposure light source is placed above the reticle 1, and the pattern on the reticle 1 is reduced by a projection lens 2 and projected onto the wafer 3, similar to the conventional apparatus. A reticle alignment mark 4 is provided around the periphery of the reticle 1 to detect the relative position of the reticle 1 and the chip 33 on the wafer 3. This reticle alignment mark 4 (hereinafter simply referred to as reticle mark 4) is sandwiched between two elongated slit-shaped chrome surfaces 4a and two chrome surfaces 4a, as shown in a circle 101 in the figure. Since the scanning light of the wafer alignment mark (to be described below and the diffracted light returning light) passes through the central part 4b, the central part 4b needs to be blank, that is, without chrome. The wafer alignment mark 17 is
For example, each chip 3 on the wafer 3 is
3. Note that since the reticle 1 and the wafer 3 are in an optically conjugate positional relationship via the projection lens 2, a wafer alignment mark (hereinafter simply referred to as wafer mark) of the chip 33 is located at the center 4b of the reticle mark 4. )
17 is imaged. The cell mark 4 and the wafer mark 17 are detected by the scanning of the He-Cd laser beam described above. The He-Cd laser light source 18 passes through a collimator lens system 19.21 and a fixed mirror 20 to a vibrating mirror 22 for beam scanning. , reach. The cross section of the laser beam is elongated into a strip-like shape by the collimator lens system. The vibrating mirror 22 is vibrating within the plane of the drawing and scans the laser beam. The scanned laser beam is guided from above the reticle 1 to the vicinity of the reticle mark 4 via a field stop 48 that limits the scanning range, an objective lens system 49, and mirrors 23 and 24. The laser beam that has reached the reticle 1 becomes an elongated (formed) strip-shaped spot light 30 in a direction perpendicular to the scanning direction by the vibrating mirror 22. At the same time, this strip-shaped spot light 30 When the band-shaped spot light 30 scans the reticle mark 4 as shown in the circle 101 and irradiates the chrome surface 4a, most of the light is reflected and the center part 4b is illuminated. When irradiated, the band-shaped spot light 30 passes through the projection lens 2 and reaches the wafer 3.The light reflected by the chrome surface 4a is photoelectrically detected by the photoelectric element 25 provided on the upper side of the reticle 1 (on the exposure light source side). On the other hand, the band-shaped spot light 30 that has passed through the center portion 4b is reduced by the projection lens 2 and scans the wafer mark 17 on the wafer 3 as a band-shaped spot light 30'. is the band-shaped spot light 3
An alignment mark is formed by arranging a plurality of minute line segment elements in a line in a direction perpendicular to the scanning direction of 0', that is, in the longitudinal direction of the band-shaped spot light 30', forming a kind of diffraction grating.

In this way, when the band-shaped spot light 30' illuminates the wafer mark 17, diffracted light is generated from the wafer mark 17. This situation will be explained based on FIG. The wafer mark 17 has a cross-sectional shape like a circle 102 in the figure, and minute line segment elements 31 are lined up in the longitudinal direction of the band-shaped spot light 30'. In a projection lens with an image side (wafer side) telescopic link, the light reflected from the wafer surface returns along the same optical path as the forward path. Therefore, the specularly reflected light generated by the wafer mark 17 passes through the projection lens 2 and forms an image on the center portion 4b of the reticle mark 4. Further, out of the diffracted light diffracted by the wafer mark, that which passes through the projection lens 2 forms an image on the center portion 4b of the reticle mark 4. For example, the N of the projection lens
, A, -0,35, the grating period of wafer mark 17 is 4μ
m, and the wavelength of the scanning laser beam is 441.6 nm, then diffracted light of approximately 0th order (regular reflection light), ±1st order, ±2nd order, and ±3rd order returns through the projection lens 2. On the other hand, in the area where the wafer mark 17 is not present, only the front reflected light of the laser beam, that is, the zero-order light, returns through the projection lens 2.

In FIG. 3, the circle 103 shows the state of the diffracted light on the pupil 6 of the projection lens 2, and positive and negative higher-order diffracted lights 34' and 35' appear at positions away from the θ-order light 30''. As is clear from FIG. 3, the longitudinal direction of the band-shaped spot light 30, 30' formed on the reticle 1 or wafer 3 is set to be perpendicular to the longitudinal direction of the zero-order light 30'' formed on the pupil 6. It will be done. In this way, the diffracted light that has reached the center 4b of the mark area of the reticle 1 passes through the reticle 1, enters the mirror 24 shown in FIG. 2, and follows the optical path to the laser light source 18. Therefore, in Fig. 2, from the mirror 20 to the field stop 4,
8. A half mirror or the like (not shown) is inserted into the optical system up to the mirror 24 to branch part of the signal light from the wafer. The branched signal light forms an image of the pupil 6 of the projection lens 2 on a conjugate plane 6' by the imaging objective lens 26, as shown in FIG.

A filter 27 that cuts only the 0th-order light of the diffracted light is arranged on the imaging surface 6', and a filter 27 that cuts only the 0th-order light of the diffracted light is arranged.
5# is focused by the focusing lens 28, and the photoelectric element 29
reach.

In the configuration described above, the photoelectric signals of the photoelectric element 25 that receives the reflected light from the chrome surface 4a of the reticle mark 4 and the photoelectric element 29 that receives the higher order diffracted light from the wafer mark 17 are respectively shown in FIGS. The result will be as shown in the figure. In FIG. 4, the aforementioned band-shaped spot light 30 scans the reticle mark 4 in the direction of the arrow. At this time, the photoelectric element 25 is
A signal 32a having two peaks as shown in the figure is output corresponding to the - of the chrome surface 4a. By accurately integrating the two peak points of the signal 32a, the center positions of the two chrome surfaces 4a can be accurately determined. In addition, chrome surface 4
Even if a is a low-reflection chrome surface, a similar signal can be obtained,
This does not cause an insufficient amount of reflected light, that is, an insufficient amount of light received by the photoelectric element 25.

The band-shaped spot light 30 passes through the reticle 1 as described above, becomes a band-shaped spot light 30' on the wafer, and scans the wafer mark 17. As mentioned above, the wafer mark 17 is a periodic structure in which minute line segment elements 31 are arranged in a line, as shown in FIG. The microelements are arranged so that the direction in which they are arranged is parallel to the longitudinal direction of the light strip 30', and the light strip 30' is scanned in the direction of the arrow in the figure. When the band-shaped spot light 30' scans the wafer mark 17, the photoelectric element 29 outputs a signal 32b as shown in the figure. Since the photoelectric element 29 is designed to receive diffracted light of the first order or higher, it outputs a signal with very good S/H. By accurately detecting the peak point of this signal 32b, the center position of the wafer mark 17 can be accurately determined. Finally, by aligning the wafer mark 17 with the center of the reticle mark 4, that is, the center of the center portion 4b, the relative positioning of the reticle 1 and the chip 33 (or the wafer 3) is performed. In this way, alignment can be performed in a direction perpendicular to the line direction of each alignment mark on the reticle and wafer.

Incidentally, when exposing the wafer 3, a part of the alignment optical system, for example, the mirror 24 and the photoelectric element 25, is configured to be able to move out of the exposure area as in the conventional case. This is a basic technology showing a case where the configuration of a conventional alignment system (FIG. 1) is used and a laser beam scanning type TTL alignment method is adopted.

Figures 6 and 7 are the basic technology of the present invention, and show the optical system of an alignment device that has fundamentally changed the arrangement of the conventional alignment optical system.The second basic technology is significantly different from the first basic technology. The difference lies in the shape of the reticle mark and that the laser spot light that scans the reticle mark and wafer mark is incident from below the reticle (projection lens side).

In FIGS. 6 and 7, the same members as in FIGS. 2 and 3,
and parts that operate in the same way use the same reference signal. FIG. 6 is a front view of the alignment device according to the second basic technology, and FIG. 7 is a view of FIG. 6 viewed from the right side. The reticle 1 is provided with a reticle mark 34 as shown in FIGS. 6 and 7. The reticle mark 34 consists of two elongated slit-shaped white areas (without chrome) 3 as shown in a circle 104 in FIG.
4a and the chrome surface 34 sandwiched between the two white parts 34a
The scanning light to the wafer mark and the diffracted light (return light) from the wafer mark, which will be described below, are reflected by the chrome surface 34b. Note that, unlike in FIGS. 2 and 3, the line direction of the reticle mark 34 is provided so as to be orthogonal to the peripheral direction of the reticle 1, that is, to face the center of the reticle 1.

Further, the wafer mark 17 on the chip 33 of the wafer 3 is also composed of a collection of minute line segment elements 31, as shown in the circle 102 in FIG. 6, as in FIGS. 2 and 3. However, the line direction of the wafer mark 17 is provided in the same direction as the line direction of the reticle mark 34.

The scanning of the reticle mark 34 by the laser beam is shown in FIG.
As in FIG. 3, scanning is performed by a spot light 30 in the form of an elongated slit. The band-shaped spot light 30 reaches the vibrating mirror 22 for beam scanning via the He-Cd laser light source 18, the collimating lens system 19.21, and the fixed mirror 20. In FIG. 6, the vibrating mirror 22 vibrates around a rotation axis in the plane of the paper to scan the laser beam. In other words, although the rotation axis is on the same plane as the paper, it is shown as shown in the figure for the sake of explanation. The scanned laser beam passes through a field stop located at a position optically conjugate with the reticle 1 (or wafer 3).
48, (which blocks unnecessary laser scanning areas), is guided from diagonally below the reticle 1 to the vicinity of the reticle mark 34 via the mirror 23, the objective lens system 49 for the reticle 1, and the millau 24, and is guided to the vicinity of the reticle mark 34. As 30,
Scanning is performed in a direction perpendicular to the line direction of the reticle mark 34.

When this band-shaped spot light 30 scans the reticle mark 34, the white area 3
At point 4a, the band-shaped spot light 30 is transmitted upward through the reticle 1, and a photoelectric element 2 disposed above the reticle mark 34 is detected.
Reach 5. On the other hand, at the chrome surface 34b, the reticle 1
It is reflected downward toward the projection lens 2. The band-shaped spot light 30 reflected by the chrome surface 34b is irradiated onto the wafer 3 as a band-shaped spot light 30' by the projection lens 2 as in FIGS. 2 and 3, and the wafer mark 1
Scan 7. At this time, diffracted light is generated as indicated by a circle 102 in FIG. At the pupil 6 of the projection lens 2, 0
Next diffracted light 30'' and higher order diffraction light 34', 3
5' appears in the same way as in Figures 2 and 3, and these diffracted lights are
The zero-diffracted light that is reflected by the chrome surface 34b of the reticle mark 34 and returns to the laser beam scanning optical system is transmitted to the objective lens system 4.
The beam is branched by a half mirror 58 disposed in a part of the wafer mark detection optical system.

As in the case of FIGS. 2 and 3, the illumination light (irradiation beam) N, A,
By limiting the numerical aperture (numerical aperture), that is, by limiting the light flux into a band shape on the pupil 6 of the projection lens 2, the diffracted light can be easily separated into diffracted lights of each order on the pupil 6. Therefore, the diffracted light from the half mirror 58 is separated into high-order diffracted lights 34# and 35# of the first order or higher by a filter 27 that blocks only the zero-order light, which is placed at a position conjugate with the pupil 6 of the projection lens 2. be done. These diffracted lights 34'', 35'' are focused by a lens 28 and reach a photoelectric element 29. Note that a part of the objective lens system 49 and the lens system 26 are optical systems that form an image of the pupil 6, and the image forming plane thereof is at the position of the filter 27.

In the configuration described above, the photoelectric signal of the photoelectric element 25 that receives the band-shaped spot light 30 transmitted through the white portion 34a of the reticle mark 34 is generated as shown in FIG. 9 in FIG. 8, the band-shaped spot light 30 is the reticle mark 34.
Scan in the direction of the arrow. At this time, the photoelectric element 25 outputs a signal 32a having two peaks because the amount of transmitted light is the largest when the band-shaped spot light 30 overlaps the position of the white portion 34a. This signal 32a is similar to that in FIG. The positional relationship between the photoelectric signal of the photoelectric element 29 and the band-shaped spot light 30' on the wafer 3 is the same as that in FIG. The output signals 32a and 32b of the photoelectric element 25 and the photoelectric element 29 cause the wafer mark 17 to be placed at the center of the reticle mark 34, that is, at the center of the chrome surface 34b.
Match reticle 1 and chip 33 (wafer 3)
Accurate relative positioning. Furthermore, wafer mark 1
When the band-shaped spot light 30' is irradiated to a location other than 7, only the specularly reflected light, that is, the 0th order light is generated, so the signal 32b is
No peak occurs.

In the second basic technology, the broken line in FIG. 6 is a region through which wafer exposure light (for example, g-line wavelength) passes. As is clear, since the alignment laser beam is incident on the surface of the reticle 1 from obliquely below, the only optical member that protrudes into the exposure area is the photoelectric element 25. Generally, optical systems for positioning are required to have particularly strict positioning accuracy. If the optical system for positioning is arranged as in the second basic technique, there is no need to move the optical system at all unless the position of the reticle mark 34 with respect to the reticle 1 changes. However, when exposing the wafer 3, it is necessary to move the photoelectric element 25 out of the exposure area. However, as described above, the photoelectric element 25 is not connected to the reticle mark 34.
Since the laser beam transmitted through the white portion 34a is simply received, there is no need to set the position with high accuracy. Therefore, compared to the first basic technique, the optical element within the exposure area can be repeatedly inserted into and removed from the exposure area at a higher speed.

FIG. 9 shows an optical system of an alignment device according to an embodiment of the present invention based on the above two basic technologies. This embodiment is an example in which the photoelectric element 25, which is the only optical element placed within the exposure area in the second basic technique, is also placed outside the exposure area. For this purpose, the white part 34a of the reticle mark 34 shown in the second basic technique is made into a diffraction grating shape similarly to the wafer mark 17, and the diffracted light generated by the irradiation with the band-shaped spot light 30 of the laser light is detected. That's how it is. Note that the configuration is exactly the same as the second basic technology except for the reticle mark and the arrangement of the charging element 25.

The reticle mark 60 is provided on the reticle 1, and the line direction (longitudinal direction) of the reticle mark 60 is located perpendicular to the peripheral direction of the reticle 1. FIG. 10 shows the relationship between the reticle marker 60 and the strip-shaped spot light 30, and the photoelectric signal.Similar to the second basic technology, the strip-shaped spot light 30 has an elongated slit-like cross section, and the arrows in FIG. (direction perpendicular to the longitudinal direction of the strip-shaped spotlight 30). The reticle mark 60 is a band-shaped spot light 30.
It consists of a lattice portion 60a in which a plurality of minute line segment elements are arranged in parallel in the longitudinal direction, and a ROM surface 60b which is similar to the second basic technology. The minute line segment elements are left blank without chrome. The chrome surface 60b is similar to the second basic technology,
The band-shaped spot light 30 is reflected toward the projection lens 2, and at the same time receives the diffracted light from the wafer mark and backlights it.

On the other hand, when the band-shaped spot light 30 begins to irradiate the grating portion 60a from obliquely below the reticle 1, diffracted light is generated due to the shape of the diffraction grating. Further, when the band-shaped spot light 30 completely overlaps the grating portion 60a, the amount of diffracted light becomes the largest. This diffracted light naturally includes first-order or higher-order diffraction light. The higher-order diffracted light is generated by spreading in the longitudinal direction of the grating portion 60a. That is, in FIG. 9, high-order diffracted light 62 is generated that travels outside the exposure area from the grating portion 60a. This diffracted light 6
2 is focused on a photoelectric element 63 (equivalent to the photoelectric element 25 in the previous basic technique) placed outside the exposure area, a signal 32a for reticle alignment similar to the first and second basic techniques can be obtained. .

According to this embodiment, since no alignment optical system is involved in the exposure area, there is no need to move the optical element during exposure, and the wafer can be exposed immediately after alignment. As a result, an alignment device that operates at extremely high speed can be obtained. Note that the photoelectric element 63 may be arranged above the reticle 1 and outside the exposure area to collect and receive high-order diffracted light generated from the reticle mark 60 above the reticle 1.

With reference to FIG. 11, signal processing for detecting the position of the alignment mark, which is common to the two basic techniques and the embodiments of the present invention, will be explained.

The scanning of the laser beam described above on each side is performed by driving the vibrating mirror 22 (for example, a galvanometer mirror) in the illumination optical system for alignment mark detection with a sine wave signal. Therefore, the band-shaped spot light 30 and the band-shaped spot light 30' become sinusoidal vibrations, so-called simple vibrations, but if the amplitude of the vibrations is made sufficiently large and the laser beam at the center of the vibrations is used for scanning, the time It is possible to scan approximately linearly. For this reason, as shown in FIGS. 2 and 6, a field diaphragm 48 is provided that allows only the scanning center portion of the laser beam to pass through. By performing scanning in a linear manner in this manner, there is no need to make corrections for nonlinearity in scanning position and time due to sinusoidal scanning.

In FIG. 11, FIG. 11(a) shows the scanning waveform of the laser beam, in which the horizontal axis represents time t and the vertical axis represents the scanning position, which is sinusoidal. In the figure, dotted lines A and B indicate two sandwiched reticle marks (chrome surface 4 a
The dotted line C indicates the position of the wafer mark 17 with respect to the scanning long waveform, and the dotted line C indicates the position of the center of the two sandwiched reticle marks with respect to the scanning waveform. By moving the reticle 1 or the wafer 3 so that the dotted lines C and D match, the wafer mark is aligned to the center of the reticle mark. FIG. 11(b) shows the output bias 32 of the photoelectric element 25 or photoelectric element 63 that detects the reticle mark.
Indicates a. The signal 32a is a sine wave scanning (back and forth) of the band-shaped spot light 30 of the laser light beam, as explained on each side.
Two peak levels Pa and Pb are generated alternately. FIG. 11(C) shows the output signal 32b of the photoelectric element 29 that detects the wafer mark when the wafer mark is at the dotted line C, and FIG. 11(d) shows the output signal 32b of the photoelectric element 29 when the wafer mark is at the dotted line C. The output signal 32b is shown when

The signals 32a, 32b are first processed by suitable signal processing circuitry (including, for example, a circuit for detecting the peak point of the signal).
Figure 1 (a), (f), (converted into a phantom binary signal). In Figure 11 (e), digital pulse signals Ap and Bp are generated at the peak point of the signal 32a, and Figure 11(f)
(g) shows the pulse signal Ap obtained in this manner, showing how a digital pulse signal Wp is generated at the peak point of the signal 32b.

The three signals Bp and Wp serve as alignment signals for positioning. From this pulse signal, in order to accurately detect the position of each alignment mark, the scanning amplitude of the laser beam should be set sufficiently thick (and each alignment mark should be scanned at the point where the linearity of the sine wave is highest). In this way, by scanning in a portion with good linearity, the pulse signal Ap,
The mutual time shift between Bp and Wp can be approximated in proportion to the positional difference (distance M) between the reticle mark and the wafer mark.

Therefore, the time Δ1t between the pulse signals Ap and Wp and the time Δ1 . If this is detected, the relative position of the reticle mark and wafer mark can be determined. And Δ1. By moving the reticle mark or wafer mark so that =Δt2 (by moving reticle 1 or wafer 3), alignment of the wafer and reticle is achieved. At this time, the pulse signal Wp becomes like a river in FIG.

Note that if the laser beam is vibrated with a large amplitude so that the linear approximation region of scanning is sufficiently wide, it is not necessary to align the vibration center of beam scanning with the center of the reticle mark. That is, the range of dotted lines A and B shown in FIG. 11(a) is
It suffices if it is within the linear approximation region of the sine wave.

In the detailed explanation of the first and second basic techniques and the present invention, the reticle mark was provided only at one place on the reticle 1, but the reticle mark was provided at only one place on the reticle 1 (the area where the actual circuit pattern is drawn). A plurality of similar reticle marks may be provided on the periphery of the reticle 1 and the wafer 3.
Performs positioning in two directions.

FIG. 12 shows another embodiment in which the reticle 1 is provided with reticle marks. In addition to positioning in two orthogonal directions using two alignment marks, a step-and-repeat type exposure apparatus requires relative θ rotation correction between the reticle and the wafer. For this purpose, reticle marks are preferably provided at three locations on the reticle 1. At this time, each chip on the wafer is also provided with a wafer mark corresponding to the reticle mark. Figure 12(a) is an example in which the longitudinal direction of the reticle mark Y and the reticle mark θ (for detecting θ rotation correction) are provided on the same line.
1 is an example in which the longitudinal directions of the reticle mark X and the reticle mark θ are provided on the same line. Further, as shown in FIG. 12), reticle marks may be provided so that the longitudinal direction extends radially on the diagonal of the reticle 1.

Furthermore, as another example of modification, the diffraction grating of the wafer mark or reticle mark is configured such that the minute line segment elements are inclined at, for example, 45 degrees with respect to the longitudinal direction, as shown in FIG. By doing this, the diffraction direction of the diffracted light changes,
It is also possible to distinguish it from scattered and diffracted light from surrounding circuit patterns.

In the embodiment described above, since the laser light is obtained from a He-Cd laser light source, the belt-shaped spot light 30' of the laser is generated on the wafer.
Scanning means exposing the photoresist. Therefore, when aligning the wafer mark on the wafer and the reticle mark on the reticle, it is desirable that the wafer mark exist within the scanning area of the band-shaped spot light 30'.

In other words, the reticle and wafer must be pre-aligned to such an extent that their respective alignment marks are roughly aligned. For this purpose, for example, another alignment optical means may be provided outside the projection lens, and the relative positions of the wafer and reticle may be adjusted by detecting specific marks on the wafer.

(Off-Axis Alignment) Generally, a wafer is exposed multiple times with a simple reticle pattern to produce one integrated circuit chip. Therefore,
Each time, as in the embodiment of the present invention, TTL, step,
When alignment is performed, in the case of a positive resist, wafer marks are removed by development and etching processing. Therefore, during the first exposure of the wafer, a plurality of wafer marks are printed together in one chip.

Reticle marks are provided on each of the plurality of reticles at different positions corresponding to the wafer marks.

Note that the alignment mark may be one on the reticle and a pair of sandwich marks on the wafer, or the same processing can be performed even if a large number of alignment marks are arranged.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing a positioning device according to the prior art, FIGS. 2 and 3 are diagrams showing an optical system of the positioning device according to the first technology, which is the basis of the present invention, and FIG. FIG. 5 is a diagram showing photoelectric signals from the reticle marks shown in FIGS. 2 and 3. FIG. 5 is a diagram showing photoelectric signals from the wafer marks shown in FIGS. 2 and 3. FIGS. 6 and 7 8 is a diagram showing the optical system of the alignment device according to the second technique that is the basis of the present invention, FIG. 8 is a diagram showing photoelectric signals from the reticle marks shown in FIGS. 6 and 7, and FIG. , FIG. 10 is a diagram showing a photoelectric signal from the reticle mark of the embodiment shown in FIG. 9, and FIG. 11 is applied to the embodiment of the present invention. Figure 12 is a diagram showing the layout of the reticle mark; Figure 13 is a diagram showing an example of changing the shape of the diffraction grating of the wafer mark or reticle mark. It is. [Explanation of symbols of main parts] Transcript・・−・−・−−１−−・・−１−−−−−・−
・−・・−・−・1 First mark area・−・・・・・−
4 Transferred object ---・----------------- 3 Second mark area ---- 17 Irradiation means ---
・・・・・・・−・−・・・・−18 Applicant: Nippon Kogaku Kogyo Co., Ltd. Figure 2 Figure 3 Figure 6 Figure 7 Figure 4 Figure 5 Figure 11 Figure 12 (( 1) (b) (C) Figure 7

Claims (2)

[Claims] (1) When projecting a pattern on a mask onto an object to be transferred via a projection optical system, optical information from a first mark area provided around the pattern on the mask is detected, and light information provided on the object to be transferred is detected. In the apparatus for relatively aligning the mask and the transferred object by detecting optical information from the second mark area via the projection optical system, the optical information is detected between the mask and the projection optical system. An irradiation means is provided for irradiating coherent illumination light toward the first mark region from outside the projection optical path of the pattern of the mask; the first mark region is irradiated with diffracted light directed outside the projection optical path by the irradiation of the illumination light. A positioning apparatus comprising: a diffraction pattern section that generates a light beam; and a light reflection section that directs the illumination light toward the second mark area via the projection optical system. (2) The diffraction pattern portions of the first mark area are formed at two locations sandwiching the light reflection portion, and pass through the center of the light reflection portion, and the diffraction pattern portions of the first mark region are formed at two locations sandwiching the light reflection portion, and pass through the center of the line symmetry of the diffraction pattern portions of the two diffraction locations. 2. The apparatus according to claim 1, wherein the direction of the line is directed approximately toward the center of the mask.